Image registration method and device based on point cloud data

Abstract

The invention discloses an image registration method based on point cloud data. The method comprises the steps: griding a CAD model into a triangular patch model; calculating local features and spatial positions of all model point pairs in the triangular patch model, and establishing a local feature model; acquiring point cloud data of a to-be-measured workpiece; calculating local features of actual point pairs; matching the local features of the actual point pairs with the local features in a local feature model to obtain a coarse registration rigid transformation matrix; calculating a fine registration rigid transformation matrix of the CAD model and the point cloud data based on the coarse registration rigid transformation matrix; and performing rotation and translation transformation on the point cloud data according to the precise registration rigid transformation matrix, so that the CAD model is matched with the point cloud data of the to-be-detected workpiece. Through the method, the local features of the to-be-measured workpiece can be accurately calculated, when the accuracy of the local features is high, the precision of coarse registration can be improved, and meanwhile, an accurate initial transformation matrix can be provided for fine registration. The invention further provides an image registration device based on the point cloud data.

Description

technical field [0001] The present application relates to the technical field of image processing, in particular to an image registration method and device based on point cloud data. Background technique [0002] The surface defects of industrial products have a negative impact on the aesthetics, comfort and performance of the products, so manufacturers detect the surface defects of workpieces in order to find and control them in time. In the process of surface defect detection, the CAD standard model of the workpiece to be tested is usually matched with the point cloud data of the workpiece to be tested in real time, that is, the image registration method, and then the workpiece to be tested is detected for defects. Image registration methods generally include coarse registration and fine registration. Among them, the coarse registration is to roughly unify the point clouds in different coordinate systems into a global coordinate system for initial coordinate transformatio...

AI Technical Summary

Problems solved by technology

There is a certain randomness when selecting feature points in the standard point cloud or the point cloud of the workpiece to be tested, which makes the rough registration rigid transformation matrix obtained in the rough registration process also have a certain randomness, resulting in a lower accuracy of the rough registration results. Low, which in turn increases the mis-match probability of fine registration

At present, by increasing the number of iterations in the rough registration process, the accuracy of the rough registration results can be improved, but the increase in the number of iterations has caused the problems of long matching time and low matching efficiency.

Images